RVF Data Processing Pipeline - Main Script

Overview

The targets contained in the predictor_data_processing_targets.R script integrate environmental data from the Africa Environmental Data Pipeline with Rift Valley fever (RVF) outbreak records from the WAHIS (World Animal Health Information System) database to create a unified dataset for epidemiological modeling.

Pipeline Components

1. Data Import

• Continental Boundaries: Creates Africa polygon and raster template (0.1° resolution)
• Base Predictors: Downloads processed environmental data from AWS
• RVF Outbreaks: Retrieves RVF outbreak records from WAHIS
• Response Variable: Generates RVF response data based on outbreak records

4. Data Integration

• Joins RVF response data with environmental predictors
• Creates unified dataset for modeling

5. Data aggregation

(In progress - missing lagged predictors)

• Imports administrative boundaries (e.g., South Africa districts)
• Aggregates pixel-level data to administrative units
• Outputs district-level datasets for analysis
• Lags important predictors

Key Features

• Multi-project Architecture: Uses targets projects to separate data acquisition from analysis
• CAPSULE Support: Manages R package dependencies for reproducibility
• AWS Integration: Syncs data with cloud storage
• Modular Design: Separate sections for different processing stages
• Error Handling: Continues execution even if some components fail

Output

The pipeline produces district-level ready for epidemiological modeling datasets by:

• Fetching environmental predictors
• Adding RVF outbreak information
• Aggregating data to administrative boundaries
• Lagging importat predictor variables